tomaarsen/reranker-modernbert-base-msmarco-bce

CrossEncoder based on answerdotai/ModernBERT-base

This is a Cross Encoder model finetuned from answerdotai/ModernBERT-base using the sentence-transformers library. It computes scores for pairs of texts, which can be used for semantic textual similarity, semantic search, paraphrase mining, text classification, clustering, and more.

Model Details

Model Description

• —Model Type: Cross Encoder
• —Base model: answerdotai/ModernBERT-base <!-- at revision 8949b909ec900327062f0ebf497f51aef5e6f0c8 -->
• —Maximum Sequence Length: 8192 tokens
• —Number of Output Labels: 1 label <!-- - Training Dataset: Unknown --> <!-- - Language: Unknown --> <!-- - License: Unknown -->

Model Sources

• —Documentation: Sentence Transformers Documentation
• —Documentation: Cross Encoder Documentation
• —Repository: Sentence Transformers on GitHub
• —Hugging Face: Cross Encoders on Hugging Face

Usage

Direct Usage (Sentence Transformers)

First install the Sentence Transformers library:

pip install -U sentence-transformers

Then you can load this model and run inference.

from sentence_transformers import CrossEncoder

# Download from the 🤗 Hub
model = CrossEncoder("tomaarsen/reranker-modernbert-base-msmarco-bce")
# Get scores for pairs of texts
pairs = [
    ['what gb sp model is bright', 'From the list above it is easy to understand that Gegabyte is bigger than Megabyte. Or GB is bigger between MB and GB. Thanks.'],
    ['does immunotherapy work', 'The US and Drug Administration (FDA) this week convened a panel of outside experts to weigh in on the readiness of a first-of-its-kind cancer therapy. The treatment, which works by tweaking a patientâ\x80\x99s own cells, is a type of immunotherapy called CAR T-cell therapy and has been in clinical trials for several years. One drug maker is now seeking FDA approval to use the treatment in pediatric and young adult patients ages 3 to 25 with B-cell acute lymphoblastic leukemia (ALL) that has not responded to standard care.'],
    ['how long to wear oasis contacts', 'There is something wrong with my Xperia Z last week, and I did a resetting to make the phone become original, but I forgot backing up some important contacts, so I have to find the way for Sony Xperia Z contacts recovery, finanlly, I got this Android Data Recovery software to recover lost contacts from my phone.here is something wrong with my Xperia Z last week, and I did a resetting to make the phone become original, but I forgot backing up some important contacts, so I have to find the way for Sony Xperia Z contacts recovery, finanlly, I got this Android Data Recovery software to recover lost contacts from my phone.'],
    ['water baby definition', "Someone very comfortable in the water, Good swimmers, and never scared while in bodies of water. There's Jadine, back in the lake. She's such a water baby. #water #aqua #babies #water babies #water kids."],
    ['youngest suicide case', 'Samantha Kuberskki was found hanging by a belt at her home in Oregon after being sent to her room for arguing with her mother. A six-year-old girl who was sent to her room for punishment is feared to be one of the youngest people to have ever committed suicide in the U.S. Samantha Kuberskki was found hanging by a belt at her home in Oregon after being sent to her room for arguing with her mother. Her death was ruled as suicide by the coroner - sparking a bitter row with police who investigated her death and insist it was an accident.'],
]
scores = model.predict(pairs)
print(scores.shape)
# (5,)

# Or rank different texts based on similarity to a single text
ranks = model.rank(
    'what gb sp model is bright',
    [
        'From the list above it is easy to understand that Gegabyte is bigger than Megabyte. Or GB is bigger between MB and GB. Thanks.',
        'The US and Drug Administration (FDA) this week convened a panel of outside experts to weigh in on the readiness of a first-of-its-kind cancer therapy. The treatment, which works by tweaking a patientâ\x80\x99s own cells, is a type of immunotherapy called CAR T-cell therapy and has been in clinical trials for several years. One drug maker is now seeking FDA approval to use the treatment in pediatric and young adult patients ages 3 to 25 with B-cell acute lymphoblastic leukemia (ALL) that has not responded to standard care.',
        'There is something wrong with my Xperia Z last week, and I did a resetting to make the phone become original, but I forgot backing up some important contacts, so I have to find the way for Sony Xperia Z contacts recovery, finanlly, I got this Android Data Recovery software to recover lost contacts from my phone.here is something wrong with my Xperia Z last week, and I did a resetting to make the phone become original, but I forgot backing up some important contacts, so I have to find the way for Sony Xperia Z contacts recovery, finanlly, I got this Android Data Recovery software to recover lost contacts from my phone.',
        "Someone very comfortable in the water, Good swimmers, and never scared while in bodies of water. There's Jadine, back in the lake. She's such a water baby. #water #aqua #babies #water babies #water kids.",
        'Samantha Kuberskki was found hanging by a belt at her home in Oregon after being sent to her room for arguing with her mother. A six-year-old girl who was sent to her room for punishment is feared to be one of the youngest people to have ever committed suicide in the U.S. Samantha Kuberskki was found hanging by a belt at her home in Oregon after being sent to her room for arguing with her mother. Her death was ruled as suicide by the coroner - sparking a bitter row with police who investigated her death and insist it was an accident.',
    ]
)
# [{'corpus_id': ..., 'score': ...}, {'corpus_id': ..., 'score': ...}, ...]

<!--

Direct Usage (Transformers)

<details><summary>Click to see the direct usage in Transformers</summary>

</details> -->

<!--

Downstream Usage (Sentence Transformers)

You can finetune this model on your own dataset.

<details><summary>Click to expand</summary>

</details> -->

<!--

Out-of-Scope Use

List how the model may foreseeably be misused and address what users ought not to do with the model. -->

Evaluation

Metrics

Cross Encoder Reranking

• —Datasets: NanoMSMARCO, NanoNFCorpus and NanoNQ
• —Evaluated with <code>CERerankingEvaluator</code>

Cross Encoder Nano BEIR

• —Dataset: NanoBEIR_mean
• —Evaluated with <code>CENanoBEIREvaluator</code>

<!--

Bias, Risks and Limitations

What are the known or foreseeable issues stemming from this model? You could also flag here known failure cases or weaknesses of the model. -->

<!--

Recommendations

What are recommendations with respect to the foreseeable issues? For example, filtering explicit content. -->

Training Details

Training Dataset

Unnamed Dataset

• —Size: 19,990,000 training samples
• —Columns: <code>query</code>, <code>answer</code>, and <code>label</code>
• —Approximate statistics based on the first 1000 samples: | | query | answer | label | |:--------|:------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------|:------------------------------------------------| | type | string | string | int | | details | <ul><li>min: 10 characters</li><li>mean: 34.21 characters</li><li>max: 197 characters</li></ul> | <ul><li>min: 82 characters</li><li>mean: 350.38 characters</li><li>max: 860 characters</li></ul> | <ul><li>0: ~73.10%</li><li>1: ~26.90%</li></ul> |
• —Samples: | query | answer | label | |:-------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------| | <code>who plays the trickster on flash</code> | <code>The Flash (2014 TV series) The Flash is a TV show based on the fictional character Flash, a costumed superhero crime-fighter who appears in comic books published by DC Comics.</code> | <code>0</code> | | <code>what type of business is plastics engineering company</code> | <code>Plastics Engineering Company is a leading North American manufacturer of phenolic resins and thermoset molding materials, selling products under its trademark Plenco. If you have a phenolic resin or thermoset molding material project, chances are, the Plenco team can make it work. We've been doing it for over 80 years. Come and benefit from the Plenco difference. Plastics Engineering Company, a family owned and managed business founded in 1934, established as its corporate mission a sincere desire to respond efficiently to the needs of our customers through development, manufacture, and servicing of useful, high-value products.</code> | <code>1</code> | | <code>what is allianz global assistance</code> | <code>Please choose 'Allianz Direct Customers' for Car, Home, Pet, Boat and Horse & Rider Insurance. Allianz Direct Customers Allianz Direct Customers Car, Home, Pet, Boat and Horse & Rider Insurance. Phone. In the Republic of Ireland: 01 448 48 48. Outside Republic of Ireland: 00 353 1 448 48 48. Opening Hours: Monday to Friday 8am - 6pm and Saturday 9am - 1pm.</code> | <code>0</code> |
• —Loss: <code>BinaryCrossEntropyLoss</code> with these parameters:

  {
      "activation_fct": "Identity",
      "pos_weight": 4
  }

Evaluation Dataset

Unnamed Dataset

• —Size: 10,000 evaluation samples
• —Columns: <code>query</code>, <code>answer</code>, and <code>label</code>
• —Approximate statistics based on the first 1000 samples: | | query | answer | label | |:--------|:------------------------------------------------------------------------------------------------|:-------------------------------------------------------------------------------------------------|:------------------------------------------------| | type | string | string | int | | details | <ul><li>min: 11 characters</li><li>mean: 33.77 characters</li><li>max: 215 characters</li></ul> | <ul><li>min: 73 characters</li><li>mean: 351.17 characters</li><li>max: 935 characters</li></ul> | <ul><li>0: ~75.80%</li><li>1: ~24.20%</li></ul> |
• —Samples: | query | answer | label | |:---------------------------------------------|:---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------|:---------------| | <code>what gb sp model is bright</code> | <code>From the list above it is easy to understand that Gegabyte is bigger than Megabyte. Or GB is bigger between MB and GB. Thanks.</code> | <code>0</code> | | <code>does immunotherapy work</code> | <code>The US and Drug Administration (FDA) this week convened a panel of outside experts to weigh in on the readiness of a first-of-its-kind cancer therapy. The treatment, which works by tweaking a patient’s own cells, is a type of immunotherapy called CAR T-cell therapy and has been in clinical trials for several years. One drug maker is now seeking FDA approval to use the treatment in pediatric and young adult patients ages 3 to 25 with B-cell acute lymphoblastic leukemia (ALL) that has not responded to standard care.</code> | <code>0</code> | | <code>how long to wear oasis contacts</code> | <code>There is something wrong with my Xperia Z last week, and I did a resetting to make the phone become original, but I forgot backing up some important contacts, so I have to find the way for Sony Xperia Z contacts recovery, finanlly, I got this Android Data Recovery software to recover lost contacts from my phone.here is something wrong with my Xperia Z last week, and I did a resetting to make the phone become original, but I forgot backing up some important contacts, so I have to find the way for Sony Xperia Z contacts recovery, finanlly, I got this Android Data Recovery software to recover lost contacts from my phone.</code> | <code>0</code> |
• —Loss: <code>BinaryCrossEntropyLoss</code> with these parameters:

  {
      "activation_fct": "Identity",
      "pos_weight": 4
  }

Training Hyperparameters

Non-Default Hyperparameters

• —eval_strategy: steps
• —per_device_train_batch_size: 128
• —per_device_eval_batch_size: 128
• —learning_rate: 8e-05
• —num_train_epochs: 1
• —warmup_ratio: 0.1
• —seed: 12
• —bf16: True
• —dataloader_num_workers: 4
• —load_best_model_at_end: True

All Hyperparameters

<details><summary>Click to expand</summary>

• —overwrite_output_dir: False
• —do_predict: False
• —eval_strategy: steps
• —prediction_loss_only: True
• —per_device_train_batch_size: 128
• —per_device_eval_batch_size: 128
• —per_gpu_train_batch_size: None
• —per_gpu_eval_batch_size: None
• —gradient_accumulation_steps: 1
• —eval_accumulation_steps: None
• —torch_empty_cache_steps: None
• —learning_rate: 8e-05
• —weight_decay: 0.0
• —adam_beta1: 0.9
• —adam_beta2: 0.999
• —adam_epsilon: 1e-08
• —max_grad_norm: 1.0
• —num_train_epochs: 1
• —max_steps: -1
• —lr_scheduler_type: linear
• —lr_scheduler_kwargs: {}
• —warmup_ratio: 0.1
• —warmup_steps: 0
• —log_level: passive
• —log_level_replica: warning
• —log_on_each_node: True
• —logging_nan_inf_filter: True
• —save_safetensors: True
• —save_on_each_node: False
• —save_only_model: False
• —restore_callback_states_from_checkpoint: False
• —no_cuda: False
• —use_cpu: False
• —use_mps_device: False
• —seed: 12
• —data_seed: None
• —jit_mode_eval: False
• —use_ipex: False
• —bf16: True
• —fp16: False
• —fp16_opt_level: O1
• —half_precision_backend: auto
• —bf16_full_eval: False
• —fp16_full_eval: False
• —tf32: None
• —local_rank: 0
• —ddp_backend: None
• —tpu_num_cores: None
• —tpu_metrics_debug: False
• —debug: []
• —dataloader_drop_last: False
• —dataloader_num_workers: 4
• —dataloader_prefetch_factor: None
• —past_index: -1
• —disable_tqdm: False
• —remove_unused_columns: True
• —label_names: None
• —load_best_model_at_end: True
• —ignore_data_skip: False
• —fsdp: []
• —fsdp_min_num_params: 0
• —fsdp_config: {'minnumparams': 0, 'xla': False, 'xlafsdpv2': False, 'xlafsdpgrad_ckpt': False}
• —fsdp_transformer_layer_cls_to_wrap: None
• —accelerator_config: {'splitbatches': False, 'dispatchbatches': None, 'evenbatches': True, 'useseedablesampler': True, 'nonblocking': False, 'gradientaccumulationkwargs': None}
• —deepspeed: None
• —label_smoothing_factor: 0.0
• —optim: adamw_torch
• —optim_args: None
• —adafactor: False
• —group_by_length: False
• —length_column_name: length
• —ddp_find_unused_parameters: None
• —ddp_bucket_cap_mb: None
• —ddp_broadcast_buffers: False
• —dataloader_pin_memory: True
• —dataloader_persistent_workers: False
• —skip_memory_metrics: True
• —use_legacy_prediction_loop: False
• —push_to_hub: False
• —resume_from_checkpoint: None
• —hub_model_id: None
• —hub_strategy: every_save
• —hub_private_repo: None
• —hub_always_push: False
• —gradient_checkpointing: False
• —gradient_checkpointing_kwargs: None
• —include_inputs_for_metrics: False
• —include_for_metrics: []
• —eval_do_concat_batches: True
• —fp16_backend: auto
• —push_to_hub_model_id: None
• —push_to_hub_organization: None
• —mp_parameters:
• —auto_find_batch_size: False
• —full_determinism: False
• —torchdynamo: None
• —ray_scope: last
• —ddp_timeout: 1800
• —torch_compile: False
• —torch_compile_backend: None
• —torch_compile_mode: None
• —dispatch_batches: None
• —split_batches: None
• —include_tokens_per_second: False
• —include_num_input_tokens_seen: False
• —neftune_noise_alpha: None
• —optim_target_modules: None
• —batch_eval_metrics: False
• —eval_on_start: False
• —use_liger_kernel: False
• —eval_use_gather_object: False
• —average_tokens_across_devices: False
• —prompts: None
• —batch_sampler: batch_sampler
• —multi_dataset_batch_sampler: proportional

</details>

Training Logs

<details><summary>Click to expand</summary>

• —The bold row denotes the saved checkpoint. </details>

Framework Versions

• —Python: 3.11.10
• —Sentence Transformers: 3.5.0.dev0
• —Transformers: 4.49.0.dev0
• —PyTorch: 2.6.0.dev20241112+cu121
• —Accelerate: 1.2.0
• —Datasets: 3.2.0
• —Tokenizers: 0.21.0

Citation

BibTeX

Sentence Transformers

@inproceedings{reimers-2019-sentence-bert,
    title = "Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks",
    author = "Reimers, Nils and Gurevych, Iryna",
    booktitle = "Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing",
    month = "11",
    year = "2019",
    publisher = "Association for Computational Linguistics",
    url = "https://arxiv.org/abs/1908.10084",
}

<!--

Glossary

Clearly define terms in order to be accessible across audiences. -->

<!--

Model Card Authors

Lists the people who create the model card, providing recognition and accountability for the detailed work that goes into its construction. -->

<!--

Model Card Contact

Provides a way for people who have updates to the Model Card, suggestions, or questions, to contact the Model Card authors. -->